Recent Changes - Search:

G3 Wiki Read First

* HomePage

* What is Gauge 3?

How To Contribute Content

Contact Administrator


* Machining Steel Wheels

* Thermodynamics 201

Design Parameters

Going around the Bend...

This may seen silly but it is quite possible to build a loco that is quite incapable of taking the corners on your layout. Your track may be gauge widened by 1mm but have you considered the amount of side play that your axles will have to have? Some locos have flat wheels and no flanges to enable them to corner such as BR 9F amongst others.

The flanges act as a limit to how tight a corner you may use. This coupled with the diameter of the wheels and the spacings of the axles dictates the absolute distance between the rails. Although the G3 specification say that the track distance is 63.5mm, with a flange thickness of 1.6mm and the back to back is 58mm it should give 1.15mm clearance on either side. These dimensions shrink as the wheel corners and the flanges eventually form a chord on the outside curve...

The following table is based on a wheel of 50mm diameter with a flange of 2.3mm depth, 1.6mm thick.

Central Axle Play -No Gauge Widening Curve Radius In Metres

Axle Spacing mm0mm side play1mm side play2mm sideplay

Central Axle Play -1mm Gauge Widening Curve Radius In Metres

Axle spacing mm0mm side play1mm side play2mm side play

The axle spacings in the tables above are for symmetrical locomotives and rolling stock. Thus a 70mm axle spacing would equate to a total fixed wheelbase of 140mm. There are several examples of locomotives with asymmetrical axle spacing. In the case of a locomotive with an axle spacing of 70mm and 90mm then it is advisable to use the figures for the 90mm axle spacing.

Independent Rotating Wheels (IRW)

This technique uses wheels not fixed to their axles. Thus it is quite possible to spin the wheels in differing directions. The coning on the wheel normally supplies the cornering ability of the rigid wheel. However sometimes this is not enough and the inner and outer wheels can thus rotate at very different speeds. This enables the wheels to take very tight corners.

One member calculated the turning radius of a G3 wheel with a fixed axle and 3 degree coning. This came out to 19.5 metres -which is about 21 chains in real life. This would be a realistic curve for a full size loco. He next worked out what it would be with 1mm gauge widening and found that it would be 9.5m or 10 chains -again a realistic figure for a "tight" curve. I have curves at 3m or 3 1/4 chains radius. This technique permits carriages and wagons to take the corner. It is still possible to use standard wheel designs but there will be a great deal of friction around an overly tight radius resulting in unequal contact velocities and therefore additional drag.

This is a study paper by Siemens on the subject. Study Paper

Pony Truck and Radial Axle Calculations.

The action of a pony truck for a model is different for that of a genuine loco. For example the axle carries very little of the weight of the loco and corners are squeakily tight. If you assume 1 yard is 1 chain then the radii of a typical G3 corner of 4.5m is 4.91 chains... A real loco would find 8 chains a screech...

There are two basic formulae for calculating the distances required.

The formula worked out by von Borries.

R = (D2 - E2 ) / 2D

The formula worked out by J.D.Baldry.

R = 0.5 (E- ( F2 / E))


  • R = the radius of the radial axles, or the effective length of the bar in the case of a pony axle.
  • D = the TOTAL wheel base of the loco.
  • B = the RIGID wheel base of the loco.
  • E = the distance from the axis of the truck or radial axle to the centre of the rigid wheelbase.
  • F = (B / 2)

The minimum total lateral movement G may be found, (using the von Borres equation), for any given curve of radius by:

G = (D2 - B2 ) / 2R

DIY Design Considerations...

Well, having looked at the published designs and found that there is nothing that can be adapted into what you wish to build you are left with one option -that is DIY... The following procedures will help you on the path to a successful design.

Powering the wheels, Purism v Cheating?

Purism is classed as making it exactly as the original was -whether this be steam, electric, or diesel. Cheating is classed as making it move by A.N.Other means... Purism will require access to the original drawings -or something “generic” that can be used in its place. Cheating will involve the use of substitute power sources. A prime example of a Purism v Cheating situation is the Gauge 1 “Project” loco. It is a steam locomotive using a single cylinder with a slip eccentric. However there are several adaptations that use dummy cylinders on the outside. Thus the loco is powered by steam -the Cheating being the internal single cylinder.

Henry Greenly both designed model steam locomotives and an electric motor designed to power steam outline locos -this fitted inside the boiler and was called (not surprisingly) a “Boiler Motor”. The concept of the “Powered Tender” (an anathema to some builders) was also developed by him.

Materials, the choices.

The Modeller will reach for Wood and Plastic -The Engineer for Steel and Brass. The time is now to decide which parts of your model should be “engineered” and which parts should be “modelled”. Normally any parts that are hot are metal as are any parts that have to take wear. Having decided what will power your model (Steam, Electric, Diesel) you now should decide what materials are to be used. The materials you use will also be decided by the tools that you have.

  • For making frames and bogies there is very little to beat sheet brass -but it is expensive.
  • Sheet steel is cheaper -but more difficult to work.
  • Tinplate and galvanized sheet are available -but have their own unique soldering problems. Rather vile acid based fluxes and subsequent water washing and neutralising are required. Rivets and soldering or brazing would be the techniques used here.
  • Plastic sheets (ABS, Plasticard etc) can be worked via “crack and snap” techniques quite easily up to 60 thous thickness after that some form of folding break is required -or sawing.
  • Acrylic and Polycarbonate would have to be hand sawn at slow speeds, preferably with a wet dripping sponge to keep the saw cut cool.
  • The use of Cyano Acrylates and Epoxies along with “solvent welding” techniques would be used here.

Finally there is wood...

Several locomotives have had wooden cabs and wooden bodywork (Swedish electric locos are an example). Lime, AKA Basswood, is normally the choice here. Although Obeche and Jellotong are extremely good -the price reflects this... Japanese wood working tools are commonly used as they leave a straight “polished” cut and do not rip the grain ends. The glues used would be Epoxies, Resins and Aliphatic wood glues. The use of “White PVA” should be avoided as not all are waterproof!

Cornering, the Model chassis versus the Prototype chassis.

If you study the tables above you will see the correlation between the amount of central axle play and the amount of gauge widening required to get a loco around a bend. At most this would have been 5mm prototypically...

If we take the case of an 0-6-0 loco with an axle spacing of 70mm, we get an initial curve of 7.538m -which equates to 8 1⁄2 chains radius. This is a fairly typical radius for a prototype 0-6-0, but most of us could never use this in our gardens... It would be far better to have loco that could take flange squealingly tight curves than to have one that could never be run. Use the tables there to design your chassis to take the tightest possible curves that you think it will ever see. It might be better to even out the axle spacing on very long locos or re-jig the spacing to provide a more smoother transition into the corner. This is an example of evening out a design, in this case an express loco, to produce an improved cornering ability.

The axle spacing is now at regular 12cm intervals -the amount of axle displacement was at most 3mm. The cornering ability of the chassis is now far more easily calculated and it will enter and leave the curve more smoothly. Also do not be afraid to “junk” some of the original design to produce a more usable model chassis design. Another example of this is in the diagram above, the Helmholz truck pivots at the 50% axle spacing position rather than at the point directly above the axle (i.e. 100%). This gives a more flexible loco. Another design item to alter is the use of a Bissel truck into a free floating bogie centred with springs on an arm. This will enable the bogie to take far tighter curves whilst still acting as a guide for the loco.

Sometimes however no matter how much you try, your only solution -is to make it bend in the middle... This is my design for the bogie of a Penn RR GG1, (something I still intend to make!). The original was a 2-C0-C0-2 design and it will have to go around 3m radius curves if it is ever to sit on my track.

Suspension systems.

Another problem you are likely to encounter is that of "suspension". The classical method uses coil springs and the wheels move up and down in their hornguides. This works very well but does require some searching for the correct coiled springs to fit in the small spaces.

However there are other design alternatives...

Most people like to think that their track is razor flat -but we all know that it is not! The NYC employed what would be nowadays called a "McPherson Strut" on some of its electric locomotives. In that it had an external coil spring and damper that pushed down onto the axle.I have an NYC "S" motor using this system and it is stable at speeds over 2metres per second. The system known as "Compensation" tends to confuse people as it is similar looking to "Equalisation".

Compensation relies on NO springs with a pivoted balance bar between the two axles thus the downward force is spread between the axles depending on their distance from the pivot point.

Equalisation uses springs connected to a balance bar, but there is NO pivot point -thus the balance bar moves in a non radial motion.

Both techniques have been used quite successfully in real life with Russian and Austro Hungarian designed locomotives being great users. Some Russian designs had moveable pivots depending on the track route that the loco was to use... The Compensation method was used by the NORD and KPeV for high speed running, whilst Equalisation was used by the LMS for the 10,000 D/E loco.

Finally -there is RUBBER! I have used with some success very dense foam rubber instead of springs. This can only take low axle loads and you will need to cut a triangular section for the suspension part.

Flexichas by Mike Sharman -getting uncommon.


Everybody gets to this point! Your design seems destined to bite you on the backside no matter what you have tried. But there are several things you may not of heard of and over the years they have been forgotten as approaches...

Loading gauge?

There are several standards They can be found in the article below.

How do I get it to corner?

Diesel and Electric chassis?

The following articles make for good reading as they show the problems and solutions adopted by previous designers...

Edit - History - Print - Recent Changes - Search
Page last modified on March 19, 2018, at 09:00 AM